# Validation
Lighthouse allows you to use Laravel's validation (opens new window) for your queries and mutations.
# Single Arguments
The simplest way to leverage the built-in validation rules is to use the @rules directive.
type Mutation {
createUser(email: String @rules(apply: ["email"])): User
}
In the case of a validation error, Lighthouse will abort execution and return the validation messages as part of the response.
mutation {
createUser(email: "foobar") {
id
}
}
{
"data": {
"createUser": null
},
"errors": [
{
"message": "Validation failed for the field [createUser].",
"locations": [
{
"line": 2,
"column": 13
}
],
"extensions": {
"validation": {
"email": ["The email field must be a valid email."]
}
}
}
]
}
# Custom Error Messages
You can customize the error message for a particular argument.
@rules(
apply: ["max:280"],
messages: [
{
rule: "max"
message: "Tweets have a limit of 280 characters"
}
]
)
# Custom Validation Attribute
You can customize the attribute for a validation message.
type Mutation {
createUser(
email: String @rules(apply: ["email"], attribute: "email address")
): User
}
# Custom Validation Rules
Reference custom validation rules by their fully qualified class name.
@rules(apply: ["App\\Rules\\MyCustomRule"])
# Validating Input Objects
Rules can be defined upon Input Object Values.
input CreatePostInput {
title: String @rules(apply: ["required"])
content: String @rules(apply: ["min:50", "max:150"])
}
Using the unique
(opens new window)
validation rule can be a bit tricky.
If the argument is nested within an input object, the argument path will not match the column name, so you have to specify the column name explicitly.
input CreateUserInput {
email: String @rules(apply: ["unique:users,email_address"])
}
# Validating Arrays
When you are passing in an array as an argument to a field, you might want to apply some validation on the array itself, using @rulesForArray
type Mutation {
makeIcecream(
"You may add up to three toppings to your icecream."
topping: [Topping!] @rulesForArray(apply: ["max:3"])
): Icecream
}
You can also combine this with @rules to validate both size and contents of an argument array. For example, you might require a list of at least 3 valid emails to be passed.
type Mutation {
attachEmails(
email: [String!]! @rules(apply: ["email"]) @rulesForArray(apply: ["min:3"])
): File
}
# Validator Classes
In cases where your validation becomes too complex and demanding, you want to have the power of PHP to perform complex validation. For example, accessing existing data in the database or validating the combination of input values cannot be achieved with the examples above. This is where validator classes come into play.
Validator classes can be reused on field definitions or input types within your schema. Use the @validator directive:
input UpdateUserInput @validator {
id: ID
name: String
}
We need to back that with a validator class. Lighthouse uses a simple naming convention for validator classes,
just use the name of the input type and append Validator
:
php artisan lighthouse:validator UpdateUserInputValidator
The resulting class will be placed in your configured validator namespace. Let's go ahead and define the validation rules for the input:
namespace App\GraphQL\Validators;
use Illuminate\Validation\Rule;
use Nuwave\Lighthouse\Validation\Validator;
class UpdateUserInputValidator extends Validator
{
public function rules(): array
{
return [
'id' => [
'required'
],
'name' => [
'sometimes',
Rule::unique('users', 'name')->ignore($this->arg('id'), 'id'),
],
];
}
}
Note that this gives you access to all kinds of programmatic validation rules that Laravel provides. This can give you additional flexibility when you need it.
# Custom Validator Messages
You can customize the messages for the given rules by implementing the messages
function:
public function messages(): array
{
return [
'name.unique' => 'The chosen username is not available',
];
}
# Validator For Fields
The @validator
directive can also be used upon fields:
type Mutation {
updateUser(id: ID!, name: String): User @validator
}
In that case, Lighthouse will look for a validator class in a sub-namespace matching the parent type, in this case
that would be Mutation
, so the default FQCN would be App\GraphQL\Validators\Mutation\UpdateUserValidator
.
# Validator For Nested Inputs
Use multiple validators for complex inputs that include nested input objects. This ensures they do not grow too complex and can be composed freely.
# Caveats
# No Mutations
Validation rules that mutate the given input are not supported:
exclude_if
exclude_unless
# References
References are resolved relative to the argument or input field that rules are defined upon:
type Mutation {
foo(bar: Int, input: FooInput): ID
}
input FooInput {
bar: Int
notBar: Int @rules(apply: ["different:bar"])
}
The following mutation would pass validation, because notBar
references the bar
field of FooInput
and thus its value 1
is compared to the value 2
- which is different:
mutation {
foo(bar: 1, input: { bar: 2, notBar: 1 })
}
# Custom Rules With References
When creating custom validation rules with references, you need to tell Lighthouse which parameters are references, so it can add the full argument path:
input FooInput {
foo: Int
bar: Int @rules(apply: ["with_reference:equal_field,0,foo"])
}
In this example, equal_field
is a custom rule that checks if the argument
is the same as the one referenced by the parameter.
The parameters to with_reference
are:
- Name of the custom rule
- Indexes of the custom rule parameter that should be treated as a reference.
Specify multiple indexes separated by
_
. - The parameters for the custom rule
If you are using custom rule classes, implement WithReferenceRule::setArgumentPath()
.
Lighthouse will call this method with the argument path leading up to the validated argument before validation runs.
# Comparisons
If you need to validate the size of an integer, you need to add the
integer
validation rule before:
type Mutation {
drinkCoffee(cups: Int! @rules(apply: ["integer", "max:3"])): Energy
}
Rules that reference other fields work strictly function as such.
For example, it is not possible to use gt
to compare against a literal value,
use min
instead:
type Mutation {
bakePizza(
dough: Int @rules(apply: ["integer", "gt:water"])
water: Int @rules(apply: ["integer", "min:2"])
): User
}
# Customize Query Validation Rules
By default, Lighthouse enables all default query validation rules from webonyx/graphql-php
.
This covers fundamental checks, e.g. queried fields match the schema, variables have values of the correct type.
If you want to add custom rules or change which ones are used, you can bind a custom implementation
of the interface \Nuwave\Lighthouse\Support\Contracts\ProvidesValidationRules
through a service provider.
use Nuwave\Lighthouse\Support\Contracts\ProvidesValidationRules;
class MyCustomRulesProvider implements ProvidesValidationRules {}
$this->app->bind(ProvidesValidationRules::class, MyCustomRulesProvider::class);